‘When dust was poured out’: Creation in Job 38.36–38

Abstract

The end of the first section of the whirlwind speech in Job 38.37b–38 reads, ‘who can tilt the waterskins of the heavens, when the dust runs into a mass and the clods stick fast together?’ (ESV). The straightforward explanation, that this refers to rainfall, cannot explain the verses in the larger context of the chapter. This article carefully reviews the words used in v. 38, in particular עפר‎ ‘dust’ and יצק‎ ‘to pour’ (here ‘to run’), and points to a parallel for דבק‎ ‘stick together’ in Job 41.15. This newly collected evidence reinforces the hypothesis of Van Wolde, that the stanza is a reference to the creation of the earth. With the suggested interpretation, the poetic structure of chapter 38 is more coherent. However, for this to work, one must either swap vv. 36 and 37 or reorder 36–38 into two tricola instead of three bicola.

1. Introduction

The first section of the whirlwind speech, in Job 38, is the long-awaited answer of God to the discussions and accusations of Job and his friends. In broad strokes, the message is that to understand Job’s suffering, one must have an understanding of the whole of creation, and that Job has ‘obscure[d] the wider purposes of God’ and has ‘misrepresent[ed] Him’ (e.g. Driver and Gray, 1921, I:325). To make his case, God reviews creation in two parts: earth (vv. 4–21) and heaven (vv. 22–38). The chapter displays an enormous metaphorical variety, where God presents himself as ‘builder, giver of fertility, judge, king, sage, and warrior’ (Perdue, 2007, 206).

The last two verses are difficult to explain in the larger context of the chapter. They read,

37: מִֽי־יְסַפֵּר שְׁחָקִים בְּחָכְמָה וְנִבְלֵי שָׁמַיִם מִי יַשְׁכִּֽיב‎

Who can count the clouds in wisdom, and the bottles of the heavens, who can overturn?

בְּצֶקֶת עָפָר לַמּוּצָק וּרְגָבִים יְדֻבָּֽקוּ‎

When dust is cast to cast iron, and the clods stick together.

Both verses are problematic, although most discussion so far has centred around v. 38. The b-colon of v. 37 is taken by the majority of commentators to mean, ‘who can make it rain?’ Some of the issues with this interpretation are discussed below. First, however, the next verse must be considered. In the commentaries, one finds two main interpretations for v. 38: the first is that this verse describes dry, parched ground (i.e. the immediate cause for rain); the second is that it describes the result of rain: dry dust clumping together to mud. Most English translations do not make a choice here. Only the NLT has ‘when the parched ground is dry and the soil has hardened into clods’. Many of the more literal translations are ambiguous and could also be read to describe what happens when it rains. Thus, one can read the NIV (‘[Who can tip over the water jars of the heavens] when the dust becomes hard and the clods of earth stick together?’) as both (1) ‘who can bring rain when the (land is so dry that) dust becomes hard and the clods of earth stick together’ and (2) ‘who can bring (might be bringing) rain when (we notice that) the dust becomes hard (as mud) and the clods of earth stick together (from the rain)’. Similar ambiguities exist in other translations (KJV, NKJV, ESV, NASB, NRSV).

I now turn to the arguments provided for either reading. Duhm (1897, 188) argues that the infinitive of יצק‎ can never mean ‘to become hard’, comparing with Job 22.16, ‘their foundation (that of wicked men) was washed (יצק‎ hophal) away’ (ESV). Accordingly, מוצק‎ must mean the not yet fully solidified metal casting. Hence, v. 38 describes the ground becoming blocked with silt by heavy rain, rather than getting hard because of a drought. Implicit here is that reading מוצק‎ as solidified metal and hence dry land does not provide a good allegory: metal casting solidifies by getting cold after having been heated, whereas land becomes dry by being heated.

An early scholar making a case for parched land is Budde (1913, 247; first edition 1896). The crux of his argument is the preposition ב. According to Budde, the infinitive with ב can only indicate the state before or at the beginning of the rain, since the effect of the rain would be indicated with ל. Indeed, Joüon and Muraoka (2011, §170j) note that ‘ב rarely has a nuance which is strictly causal’ (although they go on to cite ‘rare cases with an infinitive’). However, an infinitive with ב can be either durative or instantaneous, but not both. ¹ Therefore, with Budde’s reading, the ground must either be dry during the rain or have become dry after which the rain instantaneously followed. The first is illogical, because as soon as the rain starts the ground is not dry any more; the second is illogical as well, because land becoming dry is a process and would therefore have durative aspect.

However, Budde receives support from Driver and Gray (1921, I:336; II:311–312), who explain the odd imagery of hard metal for dry land with references to Lev. 26.19 and Deut. 28.23. In Leviticus, the Hebrews are warned by God: ‘if you will not listen to me and will not do all these commandments . . . , I will break the pride of your power, and I will make your heavens like iron and your earth like bronze’ (26.14, 19 [ESV]). In Deuteronomy, the same image is used in a similar context, but the metals are swapped: the heavens will be bronze and the earth will be iron. This indeed suggests that the comparison is not made on the basis of the colours of the metals or some other distinct feature, but on the basis of the hardness of the metals, which they have in common.

According to Driver and Gray, יצק‎ can mean ‘to become vast’, which would be attested in 1 Kgs. 22.35: ‘the blood of the wound flowed (יצק‎ qal imperfect) into (אל‎) the bottom of the chariot’ (ESV). ² However, Driver and Gray correctly note that מוצק‎ is usually something solid (it is not needed to resort to מוּצֶקֶת‎ for this, like Budde (1913, 247)).

Dhorme (1967, French original 1926) writes that ‘בְּצֶקֶת‎ denotes a circumstance concomitant with that expressed in v. 37’, but then goes on to say that ‘[t]he effect of the rain is to agglomerate the dust into a compact mass’. This is physically incorrect: during the rain, dust gathers into a muddy substance; it only hardens after the rain has stopped—but then the two events are not concomitant any more. However, Dhorme insists on reading יצק‎ as ‘to harden’, a meaning ‘appropriate to its complement למוצק‎’. This issue can be resolved by taking יצק‎ as ‘to flow’; compare the discussion of יצק‎ below.

Much of the rest of the work from the 20th century does not provide new insights into the issues at hand. Kissane (1939, 271) describes v. 38 as ‘[t]he result of rain-fall. The dust caused by the prolonged drought is again transformed into a solid mass’. Hölscher (1952, 90) translates ‘[w]enn sich die Erde zu Schlamm verfestigt, und Scholle an Scholle zusammenklebt?’, noting ‘[w]örtlich: “wenn die Erde zu Gegossenem (=Festem) sich (zusammen)gießt”’. This image is unclear: Schlamm, mud, is something more liquid than Erde, so how can the Erde verfestigen (solidify) to Schlamm? Fohrer (1963) and Hartley (1988) explain the verse as the situation before rain but do not provide new arguments. Gordis (1978) does not make a choice, but refers to the discussion.

Clines (2011, 1117) wavers between two opinions when he writes that ‘[t]he rain here is the early rain of October–November, falling after the heat of the summer has cracked open the ground . . . What is depicted is the agglomeration of the dust into a compact mass by the rain’. He makes a clear choice in the second sentence quoted here, citing several of the authors discussed above for both sides of the argument. However, it is unclear on what basis the rain must then be that of October–November. If v. 38 does not indicate parched land, it seems that the verse could describe any rain, not necessarily one falling after a long drought. ³

To summarise, in the commentaries no consensus has been reached on the meaning of Job 38.38. To do justice to the preposition in בצקת‎, the verse must be translated either in the sense that it rains when the land has become dry or that it is raining when the dusty ground is clumping together. In the literature, arguments for and against reading יצק‎ as ‘to become hard’ rather than ‘to flow’ exist, and similarly, arguments for and against reading מוצק‎ as fully or not yet fully solidified metal casting have been put forward. Two alternative interpretations, which change the meaning of the verse in more or less radical ways, are also valuable for the present discussion.

In a 1932 article, Alfrink argued that the רגבים‎ that stick together in v. 38, traditionally translated ‘clods (of earth)’, are better translated by ‘stones’. Although the argument centres around the only other occurrence of רגב‎, in Job 21.33, the author also explains how this new translation helps us to understand 38.38. According to Alfrink, the Palestinian ground consists primarily of rock, with a thin layer of fertile soil on top of a layer of gravel. When it rains, the sand and gravel mix and stick together; therefore, v. 38 describes the result of rain. However, Alfrink’s reading does mangle the Hebrew. To name a few issues: the infinitive of יצק‎ is left untranslated, מוצק‎ is given the general meaning ‘festen Masse’, and עפר ורגבים‎ is taken as a single noun phrase (the subject of ידבקו‎) although it is broken up by למוצק‎.

Most authors acknowledge the variety in subject matter in the first part of the whirlwind speech, but do not attempt to explain in detail the order in which the different topics are discussed. For instance, while Kissane (1939) pays much attention to the strophic arrangement of the chapter, to the point that he moves vv. 31–32 to its beginning to force it into an arrangement with equally long strophes, he only mentions that vv. 1–15 deal with ‘[t]he creation of the universe’ and vv. 16–38 with ‘[t]he phenomena of nature’. No attempt is made to explain the order of the strophes in these subsections.

Van Wolde (2006, 61–72) argues that the speech in Job 38 follows a structural view of the cosmos divided into six parts: three regions in the heavens and three regions below the heavens. Such a view was common in ancient Mesopotamia (Horowitz 2011). Following this reading, God discusses the earth’s surface (vv. 4–15), descending to the subterranean waters (16) and the underworld (17–21), and then going up to the wind and weather (22–30) and the stars and constellations (31–35). ⁴ Vv. 37–38 must then be about the highest heaven in which true wisdom is to be found. This leads Van Wolde to a very different reading of v. 38. According to her, it describes the creation of the earth in a way reminiscent of the way Marduk creates the earth in The Bilingual Creation of the World by Marduk. In this text, Marduk creates dirt (Akkadian eperu, a cognate of עפר‎) and pours it on a raft (Horowitz 2011, 130–131). Van Wolde then wants to read v. 38 as ‘when soil was poured into a mould, the clods stuck together’ (Van Wolde 2006, 70), noting (70–71),

The corresponding usage of the words ‘pour’ and ‘soil’ . . . express the idea that in primordial times, when the heavenly jars were poured out, the clods of soil that form the lands were glued or stuck together. Thus, the collocation ‘the jars of heaven’ in the final scene of ch. 38 describes the Higher Heaven’s primordial activities.

This reading is not free of problems either. To my knowledge, the idea that during the creation of the world the Higher Heaven rained, in order for the land to become vast, would be novel in the Ancient Near East. No such thing is suggested by The Bilingual Creation of the World by Marduk. Since earth is created in the midst of the cosmic ocean, and bounds are set to the sea, the earth’s surface is kept in place automatically. This would be the natural way to understand the situation: at the seashore one can see that the otherwise loose desert sand sticks together because of the water; no primordial rain is required. In fact, even the idea that the Higher Heaven may rain (in primordial times or later) has to be explained, since the weather would be part of the Middle (or Lower, see footnote 4) Heaven.

Also reading מוצק‎ as ‘mould’ is doubtful. As seen above, Driver and Gray already noted that מוצק‎ generally refers to the metal that is cast itself, not the mould it is cast in. The word may also be related to צוק‎ and is then used for bounds set to breadth (Job 36.16; 37.10). This meaning may be more appropriate in a cosmological context, but then it should be translated with a word more general than ‘mould’, for example, ‘bound’.

Finally, note that Van Wolde’s translation seemingly ignores the waw in ורגבים‎.

2. Contribution and organisation

This article supports the general picture as sketched by Van Wolde (2006), that Job 38.38 refers to the creation of the earth. However, it offers a more detailed analysis of the available biblical evidence and pays greater attention to the problems of the proposal. I follow Van Wolde in taking the ‘dust’ (עפר‎) in v. 38 as a kind of cosmological dust of which all of creation is made. However, I deviate in the b-part, for which Van Wolde’s translation is not satisfactory for reasons outlined above.

First, I will briefly discuss some of the most influential ancient and medieval interpretations. These support the idea that v. 38 literally describes the creation of the world. Then, the meaning of עפר‎ will be discussed in more detail. In the majority of the occurrences, the word carries connotations of creation. Finally, I will consider the various readings of בצקת‎ and מוצק‎, as well as רגבים‎.

Having established the proposed reading of v. 38, I will turn to the preceding verse. Its meaning is now unclear, and as seen above, the suggestion that it describes the primordial activities of Higher Heaven (Van Wolde 2006, 71) is unsatisfactory. I will propose two solutions, both text-critical, but only one requiring an emendation of the Masoretic text. The first solution is to move v. 37 to before v. 36. It then belongs to the preceding stanza, which has a meteorological theme. The second solution is to read vv. 36–38 as two tricola rather than three bicola. The a-part of v. 37 aligns well with v. 36. The bottles of heaven are then assumed to be jars in which the עפר‎ was stored; as such, the b-part aligns well with v. 38.

3. Ancient interpretations

Unfortunately, the Qumran fragments of Job available do not attest the end of chapter 38. However, the Old Greek translation of Job can be discussed, as can Rashi’s interpretation.

For Job 38.38a, the Old Greek translation has κέχυται δὲ ὥσπερ γῆ κονία. ⁵ This can be translated as ‘dust has been poured out like soil’ (Cox 2007), but it is unclear what this would mean: how is soil poured out? The Greek is better translated as ‘dust has been poured out as (to form) earth’, taking ὥσπερ much like Hebrew ל. The b-part of the verse is translated by the Septuagint as the rather obscure κεκόλληκα δὲ αὐτὸν ὥσπερ λίθῳ κύβον ‘I have welded it as a square block of stone’ (McCrindle 1897, 30). In several ancient religions, the cosmos was envisioned as something square or cubical. With this in mind, it should be no surprise that several ancient interpreters take the verse as a reference to creation (Ehrhardt 1945, 183–185, 188–189; on the use by Constantine of Antioch, see Laderman 2009, 124–127). ⁶

Rashi also reads the verse this way. On בצקת עפר‎, he notes, ביום שׁיצקתי עפר‎ ‘on the day that I poured out dust’; on למוצק‎, ליסוד עולם באמצעיתו‎ ‘for the foundation of the world, into its midst’. On the last part, ידובקו ורגבים‎, he comments, סביבותין לצדדין עד כי נתמלא ארכו ורחבו‎ ‘around it on the sides, to the point that its length and width were filled’. This last comment reminds us of another passage from The Bilingual Creation of the World by Marduk. In CT 13 37.31–32, Marduk ‘filled in landfill at the edge of the sea. Ea established the sea, reed thicket, and dry land’ (Horowitz 2011, 131). Perhaps the image here is that the עפר‎, dry of itself, sticks together on the sea shore (as is the case on the beach), to hold the rest of the earth in place.

4. עפר‎

Traditional translations of עפר‎ (the ‘dust’ which is ‘cast to cast iron’ in v. 38) are ‘dust’, ‘(loose) earth’, ‘dirt’, ‘soil’ and ‘ashes’ (Schwarzenbach 1954; Wanke 1976; Wächter 2001). I will argue here that while the word is used in all these contexts, it must be understood in a more abstract way, meaning something like ‘stuff’ or the smallest material units out of which something is made. This would allow us to understand the עפר‎ in Job 38.38 as the particles out of which the earth is made. Thus, the verse directly refers back to the time of creation, with which it nicely closes the first section of the whirlwind speech. Analysing all 110 occurrences of עפר‎ is outside the scope of this article, but a number of occurrences that may be better understood by means of this more abstract/particular approach will be mentioned.

First, note the well-known instances where humans are made of עפר‎: Gen. 2.7; 3.19; Ps. 104.29; Job 10.9; 34.15; Eccl. 12.7. This image leads Zevit (2013, 80) to translate עפר‎ as a ‘clod of soil’. While this meaning is applicable in these cases, it is not in many others, in particular those where people throw עפר‎ on their head (Josh. 7.6; Ezek. 27.30; Am. 2.7; Job 2.12; 16.15; 42.6; Lam. 2.10). Zevit provides his translation to avoid the concept of man being formed from dust, which is indeed somewhat awkward. But this issue is also resolved by taking עפר‎ as a more general material.

Besides humans, the earth is also made of עפר‎: God ‘enclosed the dust of the earth in a measure’ (Isa. 40.12, a passage full of cosmological terminology, ESV); wisdom was set up ‘before he (God) had made . . . the first of the dust of the world’ (Prov. 8.26 [ESV]). The latter verse is one of the two cases in the Hebrew Bible of a plural form of עפר‎, supporting a translation in the direction of ‘particles’. The other occurrence is Job 28.6 which describes minuscule particles of gold. Ecclesiastes even goes so far as to say that everything is made of עפר‎ ( 3.20). Since humans and the material world are made of עפר‎, it can have connotations of mortality and humility: Gen. 18.27 (‘I have undertaken to speak to the Lord, I who am but dust and ashes’ [ESV]); Ps. 103.14. In Job 4.18–19, God’s angels are compared to ‘those who dwell in houses of clay, whose foundation is in the dust’ (ESV)—note that the pronominal suffix on ‘foundation’ can refer back to both the houses and the mortals themselves. Perhaps also the aforementioned mourning rites where people throw עפר‎ on their head arise from this concept of mortality and humility.

Understanding עפר‎ as a kind of cosmological particle also helps to understand its usage in expressions of multitude, in Gen. 13.16; 28.14; Num. 23.10; 1 Kgs. 20.10; Zech. 9.3; Ps. 78.27; Job 27.16; 2 Chr. 1.9. It also sheds new light on some verses discussed by Ridderbos (1948), who argues that in some cases עפר‎ can mean the material the netherworld is made of (mostly in Job: 17.16; 19.25; 20.11; 21.26; 40.13; 41.25; but also in Ps. 7.6; 22.30; 30.10; Isa. 26.19; Dan. 12.2). Returning to the image of Marduk pouring dirt (eperu) on a raft to create earth (Horowitz 2011, 130–131), Ridderbos’ suggestion makes sense within the larger picture of the word sketched here. While the top layer of this dirt has become the earth’s surface in all different kinds of forms (hard metal and sticky clods), what is below the earth’s surface, that is, the netherworld, would have remained plain ‘stuff’, עפר‎. ⁷

Since things are made of עפר‎, it makes sense that when something is destroyed it returns to עפר‎. The word then indicates that nothing of the remains reminds of the original object. All of the verses where humans are made of dust, with the exception of Gen. 2.7, relate to this as well; also compare Ps. 18.43 and Job 30.19. But in general, עפר‎ remains after the destruction of anything: animals (Num. 19.17); armies (2 Kgs. 13.7; Isa. 41.2); vessels, an Asherah, and altars (2 Kgs. 23.4, 6, 12, 15); the golden calf (Deut. 9.21); buildings (Lev. 14.41, 42, 45; Isa. 25.12; ⁸ 26.5; Ezek. 26.4, 12; Ps. 102.15; Neh. 3.34; 4.4). In 1 Kgs. 18.38, in the conflict between Elijah and the prophets of Baal at the Carmel, the fire of Yahweh does not consume the dust (neither dust, nor ashes, nor sand can burn). The verse should be read as, ‘the fire of Yahweh fell and consumed the offering, the wood and the stones—the dust and the water that was in the trenches it licked up’, that is, the fire did not leave anything behind.

One might even hypothesise that the well-known image of the snake as a dust-eater (Gen. 3.14; Isa. 65.25; Mi. 7.17) yields a new metaphorical meaning. If this dust cognitively refers back to the particles of which the world is made, a dust-eater is something undermining the world order. This resolves the issue that Amzallag (2017, 777) sees, that the first part of Gen. 3.14 is realistic but the second is not. Amzallag’s own suggestion relies on reading עפר‎ as ‘ore’. The snake would be an allusion to copper which rusts and must be re-melted with ore. It is unclear why such a thing must be alluded to in Gen. 3.14. A snake as undermining the world order, however, aligns very well with v. 15, which describes how the snake will be in constant conflict with mankind. The translation of עפר‎ as ‘ore’ hinges upon Job 28.2, where iron is taken from עפר‎. But reading עפר‎ as cosmological particles brilliantly underlines the core message of the section, which marvels at the wondrous deeds that humans are capable of: out of all the עפר‎, this unspecified matter, humans are able to extract iron, something concrete and useful.

In other verses, עפר‎ is not immediately related to creation or destruction. In some cases, general ‘stuff’ or ‘dirt’ remains helpful (Gen. 26.15; Lev. 17.13; Num. 5.17; Isa. 49.23; 52.2; Ezek. 24.7; Job 7.5; 14.19; 22.24). Dirt is a small, powder-like material and can therefore easily be incorporated into the semantic field of עפר‎. The other meanings that have been proposed (‘ashes’, ‘earth’, ‘soil’) are related through the concepts of destruction and creation, but cannot be understood to be at the core of the semantics of the word.

This section is not meant to offer a conclusive study of the meaning of עפר‎, but does indicate that the word often plays in the semantic fields of creation and destruction. In particular, its usage in Job centres around this topic, perhaps with exception of the mourning rites in Job 2.12; 16.15; and 42.6. Combining this with the knowledge that in Ancient Near Eastern cosmologies עפר‎ is also used as the material of which the world is made provides us with a very readable translation of Job 38.38. The verse refers back to the creation, when the עפר‎ was used to create the earth.

5. בצקת‎, מוצק‎, and רגבים‎

It was already noted above that there is disagreement among scholars with regard to the semantic range of מוצק‎ and יצק‎. The main question is whether יצק‎, when used intransitively, means ‘to flow’ or ‘to solidify’ (or both), and as a result whether the cast metal indicated by מוצק‎ is still fluid to some extent or has solidified already. This is relevant for the traditional reading of the verse: if מוצק‎ is hard metal, the verse would describe dry land, the situation before rain; to explain the verse as describing the result of rain, מוצק‎ has to be fluid still.

This question cannot be answered conclusively by looking at יצק‎ alone. The data collected by Johnson (1990) are quite clear: the root can mean both ‘to pour’ (liquids) and ‘to cast’ (metal to something solid, like rings, supports, or a mirror). Whether the root in the context of metal casting can refer to the stage in which the metal is still fluid remains an open question. It is definitely not impossible in cases like 1 Kgs. 7.33, 46.

However, in this particular case, parallelism provides support. As in 38.38, יצק‎ stands parallel to דבק‎ in Job 41.15:

Job 41:15: מַפְּלֵי בְשָׂרֹו דָבֵקוּ יָצוּק עָלָיו בַּל־יִמֹּֽוט‎

The flakes of his (Leviathan’s) skin stick together, it is cast on him without moving.

This parallel is rarely referred to in the commentaries. Duhm (1897, 200) on 41.15 refers to 38.38 and alters the qal form to a pual. However, he then admits that the meaning of 38.38 is obscure. Driver and Gray (1921, II:341–342) and Dhorme (1967, 638) reject Duhm’s emendation. While some commentators relate דבק‎ in these two verses, it seems no one deems the repetition of the pair of roots יצק - דבק‎ worth mentioning. ⁹

This root pair gives essential information for the interpretation of 38.38. If the roots relate to the hardness of the flesh in 41.15 (especially taking ‘stone’ and ‘millstone’ in v. 16 into account as well), they must relate to the hardness of the earth in 38.38. ¹⁰ This hardness fits well with the theme of creation. In numerous places, God is said to establish the earth firmly or ‘on foundations’. Compare, for instance, Ps. 104.5: ‘he established the earth on its foundations, (so that it would be) without moving for ever’. Curiously, this verse uses יסד‎ ‘to establish’ as in Job 38.4, and בל־תמוט‎ as in Job 41.15. The root יסד‎, which is frequently used for the founding of the earth, also occurs side by side with אבן‎ in Isa. 28.16 (‘I am the one who has laid as a foundation in Zion, a stone, a tested stone’ [ESV]).

The idea to relate מוצק‎ to צוק‎, as seen in Isa. 8.23 and certainly in Job 36.16, and suggested for Job 38.38 by Delitzsch (1902, 86), who translates ‘Klumpen’, is tempting, especially when considered in conjunction with the image found in Isa. 40.12: וְכָל‎ בַּשָּׁלִשׁ עֲפַר הָאָרֶץ‎ ‘Who . . . enclosed the dust of the earth in a measure’ (ESV). Job 38.38 then describes God pouring dust through a kind of funnel to create the earth. However, on the basis of the parallel root pair in 41.15, this explanation must be rejected.

There now are two ways to understand the preposition in למוצק‎. Like in the traditional reading found in the commentaries, one can understand that the עפר‎ transforms into מוצק‎ to form the earth’s surface. As seen above, the earth’s surface is sometimes likened to metal, as in Lev. 26.19 and Deut. 28.23. The other option is to envision the עפר‎ being poured onto the מוצק‎, which is then understood to be a metal plate of some kind. This plate forms the foundation of the earth on which the earth itself, with all different types of soil, is formed of עפר‎. The latter option allows us to read בצקת‎ transitively (‘the pouring of dust’), which is the dominant usage of the root (Johnson 1990, 255). Furthermore, it provides a good parallel with The Bilingual Creation of the World by Marduk, in that the מוצק‎ stands parallel to the ‘raft’ onto which Marduk poured ‘dirt’ to create the earth (Horowitz 2011, 130–131). However, to my knowledge, the image of such a metal plate existing before creation is not attested elsewhere. At present, both options for ל seem to be maintainable.

In both cases, the רגבים‎ of the b-part can be understood as a kind of clod which somehow helps to keep the עפר‎ in place. One might hypothesise that it is a word for clods from deep in the earth. This would be consistent with the one other usage in Job 21.33, where the ‘clods of the valley are sweet to’ a deceased, an image which is perhaps slightly more understandable if the רגבים‎ are from deep in the earth, closer to the netherworld. Although this suggestion receives some support from Akkadian and Syriac, ¹¹ with only two occurrences in the Hebrew Bible, which are both questionable, this remains speculation.

6. Moving verse 37

The main issue this reading of v. 38 creates is that v. 37 now stands alone. Verse 36 anticipates the creation theme of v. 38, ¹² but v. 37 is oddly interposed. But the verse is problematic in itself as well. Upon closer inspection of the straightforward translation, as reflected in all major translations and in (1) above, there is little parallelism in this verse. While there may be parallelism between שׁחקים‎ ‘clouds’ and נבלי שׁמים‎ ‘bottles of heaven’, in combination with the verbs יספר‎ ‘who can count’ and ישׁכיב‎ ‘who can overturn’, it is unclear what such parallelism would mean. Duhm (1897, 188) suggests that God counts the clouds to see whether he has enough for rain, but also remarks that it remains a peculiar idea. Also recognising the issue, Guillaume (1964, 161–162) proposes to read יספר‎ as Arabic yusaffiru, yielding ‘who sends forth the clouds in wisdom’, an image with parallels in the Qur’ān. ¹³ This idea had already been suggested by Driver (1955, 92). While it indeed provides better parallelism, it still clearly gives v. 37 a meteorological interpretation. With this reading, v. 37 fits well with vv. 34–35, but now v. 36 is oddly interposed.

It is tempting to move v. 37 with above v. 36. This would yield the following:

37: מִֽי־יְסַפֵּר שְׁחָקִים בְּחָכְמָה וְנִבְלֵי שָׁמַיִם מִי יַשְׁכִּֽיב‎

Who can count the clouds in wisdom, and the bottles of the heavens, who can overturn?

36: מִי־שָׁת בַּטֻּחֹות חָכְמָה אֹו מִֽי־נָתַן לַשֶּׂכְוִי בִינָֽה‎

Who has put wisdom in the inner parts, or who has given the mind understanding,

38: בְּצֶקֶת עָפָר לַמּוּצָק וּרְגָבִים יְדֻבָּֽקוּ‎

when the dust was flowing/solidifying to cast iron and clods that stick together?

Verse 37 belongs to the previous stanza, vv. 34–35, which discusses authority over the weather. Guillaume and Driver’s suggestion is compatible with this reading. However, it is not necessary (and must be rejected as too far-fetched); the lack of parallelism is not insurmountable, or alternatively one may turn to one of the explanations provided by Duhm (1897) and others, such as Jenni (1968, 219) who suggests that the piel of ספר‎ indicates ‘checking whether one has enough’ rather than simply ‘counting’. Calling clouds the ‘bottles of heaven’ allows for an associative link with ‘inner parts’ in v. 36, which is a description of Higher Heaven in Mesopotamian texts (Van Wolde 2006, 70). Verses 36 and 38 form a new section, which closes the preceding chapter by stating that the wisdom in nature has been established since the time of creation. This is not an unfamiliar idea; for instance, in Prov. 8.22–31, personified wisdom boasts about having been created before the earth existed.

As usual with text emendations, there are countless ways in which the original text might have become corrupted, and it is impossible to reconstruct the correct one with certainty. Note, however, that it is possible that the exchange of vv. 36 and 37 was caused by parablepsis: they both start with מי‎. The scribe would then inadvertently have copied v. 37 before v. 36, realised the mistake, and appended v. 36. Note that the Old Greek of Job has these verses in the same order as the Masoretic text, meaning the scribal error would have to have occurred relatively early.

7. A redivision into two tricola

Can my understanding of עפר‎ and v. 38 in general be maintained without emendations of the Masoretic text? The answer is yes. By restructuring vv. 36–38 as two tricola rather than three bicola, v. 37 can be explained in context as well. The stanza then reads as

36–37a: מִי־שָׁת בַּטֻּחֹות חָכְמָה אֹו מִֽי־נָתַן לַשֶּׂכְוִי בִינָֽה מִֽי־יְסַפֵּר שְׁחָקִים בְּחָכְמָה‎

Who has put in the inner parts wisdom,

or who has given the mind understanding —

who can count the clouds in wisdom?

37b–38: וְנִבְלֵי שָׁמַיִם מִי יַשְׁכִּֽיב בְּצֶקֶת עָפָר לַמּוּצָק וּרְגָבִים יְדֻבָּֽקוּ‎

And the bottles of the heavens, who could have overturned,

when the dust was flowing/solidifying to cast iron

and the clods stick together?

While tricola are frequent in both Ugaritic and Akkadian poetry, one must approach an apparent tricolon in Hebrew with scepticism. Mowinckel (1957) held that tricola can only occur in (large sections of) psalms that are fully tricolic. However, later scholars rejected his methodology, which consisted of ‘a rewriting of the Hebrew text of passages which do not fit his scheme’ (Willis 1979, 466). Others, such as Watson (1984), find that tricola are to be expected in Hebrew poetry given their use in Ugaritic and Akkadian (Watson 1984, 177), although Loretz and Kottsieper (1987, 50) correctly reaffirm some of Mowinckel’s scepticism by stating that ‘the permissibility of asserting the presence of tricola in precisely those biblical poetic texts which contain colometric difficulties is questionable’. For a more up-to-date overview of the small amount of studies on tricola so far, see Stocks (2010, 18–22).

There is more consensus on what tricola would express. Watson (1984, 183–185) states that the tricolon ‘demarcat[es] stanzas . . . coming either at the beginning or at the end, and sometimes in both places’, and that it may ‘mark a climax’, as also noted by Gordon (1965, 133) and further investigated by Yaron (1986). ¹⁴ Stocks (2010, 244) does not find any cases of tricola opening new sections in the psalms of ascents, but can confirm the other functions postulated by Watson.

The first tricolon, in vv. 36–37a, is full of parallelism. The first two cola are both formed by a sequence of מי‎, an indicative verbal form, a prepositional phrase, and a direct object. This structure is flipped in the third colon, which puts the prepositional phrase at the end (but still starts with מי‎). By flipping the structure in the last colon, all cola end with either חכמה‎ ‘wisdom’ or בינה‎ ‘understanding’.

It is true that the third colon has a yiqtol rather than a qatal. It should be read with a modal nuance: ‘(indeed,) who can count’. The first two cola provide the proof that God can ‘count the clouds in wisdom’. This expression is still odd, but less so than when it is coupled to 37b. When appended to v. 36, it may simply mean that God is the only one who is able to oversee all of creation. While most other examples of tricola use only one conjugation (in those cola that contain a verb), it does not seem that this was vital. Willis (1979, 469) notes Ps. 18.48–49a, which has a yiqtol between two participles, and in Job 10.1, a qatal is followed by two yiqtols. Also, the examples of bicola with verbs in different conjugations abound. In general, it seems that the Hebrew poet had no problem with changing the conjugation when required for semantics. ¹⁵

As for the second tricolon, the waw at the beginning is perhaps somewhat unusual, but can be used when a verse continues a train of thought (cf., in this chapter, vv. 8, 11, 15). The train of thought followed here is that given that (only) God has established wisdom and can count the clouds, he is also the one who must have (the only one who could have) overturned the bottles of heavens at the start of creation. As discussed above, the two cola of v. 38 stand in parallel. Tricola with such an A/B/B′ structure are rather common. Watson (1984, 181) provides examples (among which is also the aforementioned Job 10.1).

The redivision of vv. 36–38 into two tricola solves the issue with v. 37. By splitting it up, it does not require a meteorological meaning, which would fit better earlier in the chapter. By aligning v. 37a with v. 36, the meaning of יספר‎ is explained in a more natural way than the suggestions of Duhm (1897, 188); moreover, recourse to Arabic as with Guillaume (1964, 161–162) and Driver (1955, 92) is not needed. The bottles of heaven in v. 37b must then be the jars that held the עפר‎ before it was poured out to create the world.

The tricola have the poetic function of marking the end of the section, but they also mark the end of the traversal of the cosmos as explained by Van Wolde (2006, 65–71), and in that sense they have a climactic function. Chapter 38 describes the earth, the waters around it, the netherworld, then moves up to the three levels of heavens. The tricola in vv. 36–38 also emphasise the climax of this overview, which describes the highest heaven.

Compared to the proposal to move v. 37 before v. 36, the suggestion to redivide the text has the obvious advantage that it does not require a textual emendation. It does not yield insurmountable problems with respect to poetic structure. However, note that the same redivision cannot be applied to the Greek translation, which replaces v. 36 with a verse about needlework skills, to which v. 37a cannot be related. Therefore, the redivision presupposes that the Hebrew must have been misunderstood already rather early.

8. Concluding remarks

I began my discussion with an overview of various interpretations of Job 38.38. The traditional understanding, that it refers to rainfall, makes it difficult to integrate the verse with the rest of the section. Rain has already been mentioned earlier in the chapter (vv. 25–28; 34). Furthermore, it is clear that in 38.39 a new section begins, so one would rather expect a proper ending here. For instance, the next section begins and ends with the hunt of animals (38.39–41; 39.29–30), forming a topical inclusio. Such a structure would be missing from the first section of the whirlwind speech. I discussed the hypothesis of Van Wolde (2006), who suggested that the stanza 38.36–38 refers to the creation of the earth. While she provided comparative evidence from the Ancient Near East, internal evidence was still lacking, and several issues with the translation were ignored.

I then turned to the internal evidence available. In particular, the words עפר‎ and יצק‎ (and by extension מוצק‎) were discussed, and to a lesser extent also דבק‎ and רגב‎. As for עפר‎, it can refer to the ‘stuff of which something is made’ in general, especially when related to the creation at the beginning of time. The root pair יצק - דבק‎ in 38.38 is most likely synonymous considering its parallel in Job 41.15 (something that up to now seems to have been ignored, even though it is also relevant to the traditional explanation of Job 38.38).Thus, v. 38 tells of the pouring of plain matter (עפר‎) on or into a solid metal object to create the earth.

Admittedly, my reading of v. 38 made it somewhat difficult to explain v. 37 in context. The article discussed two possible solutions: first, to move v. 37 to before v. 36; second, to read vv. 36–38 as two tricola rather than three bicola. Although the textual corruption that the first solution presupposes can be explained by parablepsis, the second must be preferred as it does not require an emendation. The two tricola are of the forms A/A′/A″ and A/B/B′, which are both well-attested in Biblical Hebrew poetry.

The imagery of v. 38 has a parallel in The Bilingual Creation of the World by Marduk and is supported by several ancient interpretations. Most importantly, however, the reading proposed here helps us to explain the larger poetic structure of 38.2–38. The section begins and ends with the topic of creation at large (vv. 4, 38), while in the intervening stanzas God reviews specific aspects of creation (cf., again, Van Wolde 2006, 65–72). In this way, the section forms a topical inclusio, just like the next section.